Technology ❯Artificial Intelligence ❯Chatbots ❯User Interaction
The update gives the models the ability to terminate chats in extreme abuse cases as a precaution rooted in predeployment welfare testing.
